Recently I had cause to want to back up the large number of custom Slack emoji used in a workspace I’m a part of. Turns out, Slack doesn’t provide an official way to do this. I saw plenty of blog posts and Gists on the subject, but Slack seems to have thwarted all of their methods (intentionally or otherwise). Here I’ll document the general process I used for backing up nearly a thousand custom emoji, in a way that I hope proves evergreen.¹

As a prerequisite, be sure to log into the target workspace in a browser, where we have easy access to developer features.

Grab some JSON Link to heading

Clientside, custom emoji are described using a blob of JSON metadata containing names and URLs pointing directly to the image files. All we have to do is find it.

I used the Custom Emoji page within the Workspace Settings, but any page that displays all custom emoji should do, even just opening the picker in a chat window and scrolling the list. Whatever the method, next open the browser’s developer tools to look at network requests (for Firefox, Ctrl-Shift-E will open the Network tab). Refresh the page and perform only the interactions necessary to display the custom emoji, this will hopefully limit the size of the haystack we’ll be searching.

Filter the of network requests down to GETs on JSON files, then look for ones that contain a list of entries with the data we need: emoji names, and links to image files. In my case I landed on emoji.adminList.json, which paginates items 100 at a time over multiple requests to the same path. There was a helpful paging object to keep track of which page was which.

Grab some emoji Link to heading

I wound up with a series of JSON files with the following format (extraneous objects omitted):

{
  "emoji": [
    { "name": "wow", "url": "https://emoji.slack-edge.com/<...>" }
  ]
}

Given this, I whipped up a quick Bash script leveraging the all-powerful jq and curl utilities to download each image and rename it based on the display name:

#!/bin/bash

target_path=emoji

mkdir -p "$target_path"

# Get a list of all items in the emoji list, jq -c ensures that each line is a
# single item
items=$(jq -c '.emoji[]')

rowcount=$(echo "$items" | wc -l)
rowcurrent=0
failures=0

# Set an empty IFS to prevent separating on spaces, then read -r to iterate
# over lines
echo "$items" | while IFS= read -r item; do
    # Grab name and URL for the emoji, jq -r to avoid extra quotes
    name=$(echo "$item" | jq -r '.name')
    url=$(echo "$item" | jq -r '.url')

    # Parameter expansion magic to trim everything but the extension
    extension="${url##*.}"
    target_file="$target_path/$name.$extension"

    echo "Downloading '$target_file' ($((rowcurrent + 1)) of $rowcount)..."
    curl -so "$target_file" "$url"
    if [ $? -ne 0 ]; then
        >&2 echo "Failed to download $url"
        failures=$((failures + 1))
    fi

    rowcurrent=$((rowcurrent + 1))
done

echo "Successfully downloaded $((rowcount - failures)) of $rowcount emoji."

A quick $ ./slackbak.sh < emoji.adminList.json call and I had a big stack of image files. I wouldn’t call it robust by any stretch of the word, but it works.

For future readers, there’s a good chance that you’ll be looking at a different JSON structure. The general gist is this: first grab the list of emoji. Next, for each item, extract the display name and the URL. The URL will almost definitely have a meaningless to humans ID for a filename, so use the extracted name and file extension to create a new one. Finally, download the file and you’re off to the races. Just about every language on the planet has a JSON parser and HTTP client, so don’t overthink the implementation; just use whatever you’re most comfortable with, get your emotes, and get on with your day. Good luck!